sched/cache: Keep nr_pref_llc_running in the runnable domain, to fix LLC mis-scheduling buggrafted

alb_break_llc() decides whether to break LLC preference during active
load balance. It does so by testing that every runnable fair task on the
source rq prefers its LLC:

	env->src_rq->nr_pref_llc_running == env->src_rq->cfs.h_nr_runnable

But the two counters cover different sets. nr_pref_llc_running is updated
in account_llc_enqueue()/account_llc_dequeue(), next to cfs_rq->nr_queued,
so it follows queued tasks. h_nr_runnable is updated in set_delayed()/
clear_delayed() and drops delay-dequeued tasks.

So under DELAY_DEQUEUE, a preferring task that goes to sleep stays counted
in nr_pref_llc_running while h_nr_runnable falls. The equality then breaks,
alb_break_llc() returns false, and active balance is free to pull a task
off its preferred LLC. Active balance only moves runnable tasks, and this
is the only LLC check it consults: once the stopper runs, LBF_ACTIVE_LB
skips the per-task test in can_migrate_task(). The runnable set is the one
we want.

Fix it on the counter side. A task should be counted in
nr_pref_llc_running exactly while it is both queued on its preferred LLC
(pref_llc_queued) and runnable (!sched_delayed). Define that membership
once in task_pref_llc_runnable(), and adjust the counter only through
pref_llc_running_inc()/pref_llc_running_dec() from the four sites that
change either input: account_llc_enqueue(), account_llc_dequeue(),
set_delayed() and clear_delayed(). Gating every update on the same
predicate keeps the delay, wake and dequeue paths from double-counting
or underflowing; see the comments at those sites for the ordering.

nr_llc_running and sd->llc_counts are not touched and stay on queued
semantics.

Fixes: 714059f79ff0 ("sched/cache: Handle moving single tasks to/from their preferred LLC")
Closes: https://lore.kernel.org/lkml/20260827135000.735138-1-zhanxusheng@xiaomi.com/
Reported-by: Zhan Xusheng <zhanxusheng@xiaomi.com>
Suggested-by: Chen Yu <yu.c.chen@intel.com>
Signed-off-by: Tim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Kayra Cizmeci <kayracizmeci@gmail.com>
Cc: <stable@kernel.org> # v7.2.x
Link: https://patch.msgid.link/06af61afedac32e6477f57feb4d658f6c411c3af.1790035273.git.tim.c.chen@linux.intel.com

1 files changed, 147 insertions, 0 deletions

diff --git a/net/rxrpc/sysctl.c b/net/rxrpc/sysctl.c
new file mode 100644
index 000000000..46a20cf4c
--- /dev/null
+++ b/net/rxrpc/sysctl.c
@@ -0,0 +1,147 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* sysctls for configuring RxRPC operating parameters
+ *
+ * Copyright (C) 2014 Red Hat, Inc. All Rights Reserved.
+ * Written by David Howells (dhowells@redhat.com)
+ */
+
+#include <linux/sysctl.h>
+#include <net/sock.h>
+#include <net/af_rxrpc.h>
+#include "ar-internal.h"
+
+static struct ctl_table_header *rxrpc_sysctl_reg_table;
+static const unsigned int rxrpc_rx_mtu_min = 500;
+static const unsigned int rxrpc_jumbo_max = RXRPC_MAX_NR_JUMBO;
+static const unsigned int four = 4;
+static const unsigned int max_backlog = RXRPC_BACKLOG_MAX - 1;
+static const unsigned int n_65535 = 65535;
+static const unsigned int n_max_acks = 255;
+static const unsigned long one_ms = 1;
+static const unsigned long max_ms = 1000;
+static const unsigned long one_jiffy = 1;
+static const unsigned long max_jiffies = MAX_JIFFY_OFFSET;
+#ifdef CONFIG_AF_RXRPC_INJECT_RX_DELAY
+static const unsigned long max_500 = 500;
+#endif
+
+/*
+ * RxRPC operating parameters.
+ *
+ * See Documentation/networking/rxrpc.rst and the variable definitions for more
+ * information on the individual parameters.
+ */
+static struct ctl_table rxrpc_sysctl_table[] = {
+	/* Values measured in milliseconds */
+	{
+		.procname	= "soft_ack_delay",
+		.data		= &rxrpc_soft_ack_delay,
+		.maxlen		= sizeof(unsigned long),
+		.mode		= 0644,
+		.proc_handler	= proc_doulongvec_minmax,
+		.extra1		= (void *)&one_ms,
+		.extra2		= (void *)&max_ms,
+	},
+	{
+		.procname	= "idle_ack_delay",
+		.data		= &rxrpc_idle_ack_delay,
+		.maxlen		= sizeof(unsigned long),
+		.mode		= 0644,
+		.proc_handler	= proc_doulongvec_minmax,
+		.extra1		= (void *)&one_ms,
+		.extra2		= (void *)&max_ms,
+	},
+	{
+		.procname	= "idle_conn_expiry",
+		.data		= &rxrpc_conn_idle_client_expiry,
+		.maxlen		= sizeof(unsigned long),
+		.mode		= 0644,
+		.proc_handler	= proc_doulongvec_ms_jiffies_minmax,
+		.extra1		= (void *)&one_jiffy,
+		.extra2		= (void *)&max_jiffies,
+	},
+	{
+		.procname	= "idle_conn_fast_expiry",
+		.data		= &rxrpc_conn_idle_client_fast_expiry,
+		.maxlen		= sizeof(unsigned long),
+		.mode		= 0644,
+		.proc_handler	= proc_doulongvec_ms_jiffies_minmax,
+		.extra1		= (void *)&one_jiffy,
+		.extra2		= (void *)&max_jiffies,
+	},
+
+	/* Values used in milliseconds */
+#ifdef CONFIG_AF_RXRPC_INJECT_RX_DELAY
+	{
+		.procname	= "inject_rx_delay",
+		.data		= &rxrpc_inject_rx_delay,
+		.maxlen		= sizeof(unsigned long),
+		.mode		= 0644,
+		.proc_handler	= proc_doulongvec_minmax,
+		.extra1		= (void *)SYSCTL_LONG_ZERO,
+		.extra2		= (void *)&max_500,
+	},
+#endif
+
+	/* Non-time values */
+	{
+		.procname	= "reap_client_conns",
+		.data		= &rxrpc_reap_client_connections,
+		.maxlen		= sizeof(unsigned int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= (void *)SYSCTL_ONE,
+		.extra2		= (void *)&n_65535,
+	},
+	{
+		.procname	= "max_backlog",
+		.data		= &rxrpc_max_backlog,
+		.maxlen		= sizeof(unsigned int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= (void *)&four,
+		.extra2		= (void *)&max_backlog,
+	},
+	{
+		.procname	= "rx_window_size",
+		.data		= &rxrpc_rx_window_size,
+		.maxlen		= sizeof(unsigned int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= (void *)SYSCTL_ONE,
+		.extra2		= (void *)&n_max_acks,
+	},
+	{
+		.procname	= "rx_mtu",
+		.data		= &rxrpc_rx_mtu,
+		.maxlen		= sizeof(unsigned int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= (void *)&rxrpc_rx_mtu_min,
+		.extra2		= (void *)&n_65535,
+	},
+	{
+		.procname	= "rx_jumbo_max",
+		.data		= &rxrpc_rx_jumbo_max,
+		.maxlen		= sizeof(unsigned int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= (void *)SYSCTL_ONE,
+		.extra2		= (void *)&rxrpc_jumbo_max,
+	},
+};
+
+int __init rxrpc_sysctl_init(void)
+{
+	rxrpc_sysctl_reg_table = register_net_sysctl(&init_net, "net/rxrpc",
+						     rxrpc_sysctl_table);
+	if (!rxrpc_sysctl_reg_table)
+		return -ENOMEM;
+	return 0;
+}
+
+void rxrpc_sysctl_exit(void)
+{
+	if (rxrpc_sysctl_reg_table)
+		unregister_net_sysctl_table(rxrpc_sysctl_reg_table);
+}